Social stress can lead to drug craving and relapse in cocaine-dependent (CD) individuals. In addition, CD individuals often favor drug use over social interactions. Moreover, social avoidance and lack of trust are significant obstacles to effective treatment. Currently, there are no FDA approved medications for the treatment of cocaine dependence and behavioral interventions have had limited success in sustaining abstinence. Data from human neuroimaging studies using blood oxygen level dependent (BOLD) functional magnetic resonance imaging (fMRI) suggest that top-down prefrontal cortical control of amygdala reactivity to social stimuli plays an important role in mediating emotion related behavior. Dysregulation in the functional connectivity between the prefrontal cortex and amygdala has been found in CD subjects at rest and attenuated corticolimbic functional connectivity was associated with a shorter time to relapse. Thus, dysregulation in corticolimbic network activity may underscore the vulnerability of CD individuals to social stress. In addition, interventions that restore functional connectivity between the prefrontal cortex and amygdala, and attenuate bottom-up amygdala drive may reduce anxiety and improve treatment outcomes for CD individuals. Oxytocin (OT) is an anxiolytic neuropeptide that reduces amygdala reactivity to aversive social cues. In addition, OT increases functional connectivity between the amygdala and prefrontal cortex in patients with generalized social anxiety disorder. The broad and long-term objectives of this proposal are to (1) to identify the neurobiologic mechanisms that control emotional responses to social stimuli in CD individuals and (2) use these data to facilitate the development of effective therapeutic treatments and preventative strategies for behavioral disorders and disease. To meet these objectives we propose two specific aims: Specific Aim 1: To determine the impact of cocaine dependence and oxytocin on functional connectivity between corticolimbic brain regions during acute social stress. Specific Aim 2: Use an implicit facial affect recognition paradigm to determine the impact of cocaine dependence and oxytocin on amygdala activity in response to fearful faces. The BOLD signal measured during neutral faces will be subtracted from the BOLD signal measured during fearful faces. To address the hypotheses associated with Specific Aims 1 and 2 we propose a we propose a double-blind placebo (PBO) controlled study using BOLD fMRI to measure (1) corticolimbic functional connectivity during the Montreal Imaging Stress Task (MIST) and (2) amygdala activity in response to an implicit facial affect recognition paradigm in groups of CD individuals (CD n=80) and healthy non-dependent controls (HC, n=80). Prior to the scanning session, participants will receive either intranasal OT (24 IU) or PBO spray (n=40 per treatment group). Psychophysiologic interaction (PPI) analysis using the amygdala as the seed region will be used to assess significant task (stress condition > control condition) x seed interactions. Subjective anxiety and craving data will be collected at baseline and after each run of the MIST. The order of the tasks will be counterbalanced.